WO2018098502A1 - Systèmes, procédés et dispositifs pour l'analyse à base de largeur de tracés de pics - Google Patents

Systèmes, procédés et dispositifs pour l'analyse à base de largeur de tracés de pics Download PDF

Info

Publication number
WO2018098502A1
WO2018098502A1 PCT/US2017/063536 US2017063536W WO2018098502A1 WO 2018098502 A1 WO2018098502 A1 WO 2018098502A1 US 2017063536 W US2017063536 W US 2017063536W WO 2018098502 A1 WO2018098502 A1 WO 2018098502A1
Authority
WO
WIPO (PCT)
Prior art keywords
peak
height
analyte
width
chromatographic
Prior art date
Application number
PCT/US2017/063536
Other languages
English (en)
Inventor
Purnendu Dasgupta
Akinde F. KADJO
Kannan Srinivasan
Original Assignee
Board Of Regents, The University Of Texas System
Dionex Corporation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Board Of Regents, The University Of Texas System, Dionex Corporation filed Critical Board Of Regents, The University Of Texas System
Priority to CN201780084712.8A priority Critical patent/CN110234990B/zh
Priority to EP17874213.6A priority patent/EP3545296A4/fr
Publication of WO2018098502A1 publication Critical patent/WO2018098502A1/fr

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G01N30/86—Signal analysis
    • G01N30/8624—Detection of slopes or peaks; baseline correction
    • G01N30/8631—Peaks
    • G01N30/8637—Peak shape
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01D—SEPARATION
    • B01D15/00—Separating processes involving the treatment of liquids with solid sorbents; Apparatus therefor
    • B01D15/08—Selective adsorption, e.g. chromatography
    • B01D15/10—Selective adsorption, e.g. chromatography characterised by constructional or operational features
    • B01D15/16—Selective adsorption, e.g. chromatography characterised by constructional or operational features relating to the conditioning of the fluid carrier
    • B01D15/161—Temperature conditioning
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01D—SEPARATION
    • B01D15/00—Separating processes involving the treatment of liquids with solid sorbents; Apparatus therefor
    • B01D15/08—Selective adsorption, e.g. chromatography
    • B01D15/10—Selective adsorption, e.g. chromatography characterised by constructional or operational features
    • B01D15/16—Selective adsorption, e.g. chromatography characterised by constructional or operational features relating to the conditioning of the fluid carrier
    • B01D15/163—Pressure or speed conditioning
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G01N30/04—Preparation or injection of sample to be analysed
    • G01N30/16—Injection
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G01N30/26—Conditioning of the fluid carrier; Flow patterns
    • G01N30/28—Control of physical parameters of the fluid carrier
    • G01N30/30—Control of physical parameters of the fluid carrier of temperature
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G01N30/86—Signal analysis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G01N2030/022—Column chromatography characterised by the kind of separation mechanism
    • G01N2030/025—Gas chromatography
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
    • G01N30/02—Column chromatography
    • G01N2030/022—Column chromatography characterised by the kind of separation mechanism
    • G01N2030/027—Liquid chromatography

Definitions

  • This disclosure relates generally to analytical chemistry. More specifically, this disclosure pertains to all analytical techniques that produce peak-shaped responses separated in time or space, for example flow injection analysis, capillary or microchip electrophoresis and especially chromatography techniques.
  • the science of chromatography techniques addresses the separation and analysis of chemical components in mixtures.
  • This disclosure relates to techniques for the quantitation of chromatographic peaks based on a width measurement of a peak trace, and assays of purity of a putatively pure separated band, or detection of impurities therein.
  • Typical practice involves a single standard linear regression equation covering multiple concentrations/amounts for quantitation. It is well known that while linear regression minimizes absolute errors, the relative error, often of greater importance, becomes very large at low analyte concentrations. Weighted linear regression provides a solution to this, but it is notably absent from popular chromatographic data handling software. Height is often regarded as more accurate than area, especially if peaks are not well resolved in the chromatogram. Height is less affected by asymmetry and overlap, and provides less quantitation error for peaks with limited overlap. In a survey of chromatographers, area was preferred over height for better accuracy and precision. However, poor resolution or significant peak asymmetry (the two are related: high asymmetry increases the probability of overlap) induces greater error in area-based quantitation. Both area and height are affected by detector non-linearity, and detector saturation leads to clipped peaks.
  • a method of chromatographic quantitation of an analyte comprises flowing the analyte at least at a first concentration, a second concentration, and then a third concentration into a chromatographic column; detecting the analyte at the first concentration, the second concentration, and the third concentration coming out from the chromatographic column by using a chromatographic detector; obtaining a first, second, and third signal curves from the chromatographic detector, the first, second, and third signal curves being a representation of the analyte at the first, second, and third concentrations, respectively, detected by the chromatographic detector; measuring a width of a peak in each of the first, second, and third signal curves at a plurality of peak heights; calculating a plurality of calibration equations based on the first, second, third concentrations and the measured peak widths for each of the plurality of peak heights; and identifying one of the plurality of peak heights that provides the calibration equation having a lowest error.
  • the method further comprises a suppressor coupled with the chromatographic column for receiving an output from the chromatographic column, wherein the suppressor is coupled with the chromatographic detector, such that an output from the suppressor
  • a method of detecting an impurity in chromatography comprises flowing an analyte of a sample through a chromatographic column; detecting a concentration of the analyte coming out from the chromatographic column by using a chromatographic detector; obtaining a first signal curve from the chromatographic detector, the first signal curve being a representation of the concentration of the analyte detected by the chromatographic detector; measuring a first peak width W h i at a first absolute peak height hi, a second peak width J3 ⁇ 4 at a second absolute peak height h 2 , and a third peak width J3 ⁇ 4 at a third absolute peak height h 3 of a peak in the first signal curve, wherein the first absolute peak height hi, the second absolute peak height h 2 , and the third absolute peak height h 3 are different; determining a peak shape index ratio of the sample of the peak in the first signal curve with a formula comprising and identifying a presence of the impurity in the sample where the determined
  • the method further comprises flowing the analyte of the standard sample through the chromatographic column; detecting a concentration of the analyte of the standard sample coming out from the chromatographic column by using the chromatographic detector; obtaining a second signal curve from the chromatographic detector, the second signal curve being a representation of the concentration of the analyte of the standard sample detected by the chromatographic detector; measuring the first peak width W h i at the first absolute peak height hi, the second peak width W h2 at the second absolute peak height h 2 , and the third peak width J3 ⁇ 4 at the third absolute peak height h 3 of a peak in the second signal curve, wherein the first absolute peak height hi, the second absolute peak height h 2 , and the third absolute peak height h 3 are different; and determining the peak shape index ratio of the standard sample of the peak in the second signal curve with the formula.
  • the method further comprises repeating the steps above on multiple injections of the standard sample; calculating a confidence range of the peak shape index ratio at a confidence level above 90% for the standard sample; and identifying the presence of the impurity in the sample where the determined peak shape index ratio of the sample is outside of the calculated confidence range.
  • the peak of the standard sample and the analyte peak of the sample under test have a same maximum peak height.
  • the method comprises a suppressor coupled with the chromatographic column for receiving an output from the chromatographic column, wherein the suppressor is coupled with the chromatographic detector, such that an output from the suppressor is detected by the chromatographic detector.
  • a method of detecting an impurity in chromatography comprises flowing an analyte of a sample through a chromatographic column; detecting a concentration of the analyte coming out from the chromatographic column by using a chromatographic detector; obtaining a first signal curve from the chromatographic detector, the first signal curve being a representation of the concentration of the analyte detected by the chromatographic detector; measuring a first peak width W h i at a first absolute peak height h a second peak width W h2 at a second absolute peak height h 2 , a third peak width J3 ⁇ 4 at a third absolute peak height h 3 , and a fourth peak width Wh4 at a fourth absolute peak height h 4 of a peak in the first signal curve, wherein the first absolute peak height hi, the second absolute peak height h 2 , the third absolute peak height h 3 , and the fourth absolute peak height h 4 are different; determining a peak shape index ratio of the sample of
  • the method further comprises flowing the analyte of the standard sample through the chromatographic column; detecting a concentration of the analyte of the standard sample coming out from the chromatographic column by using the chromatographic detector; obtaining a second signal curve from the chromatographic detector, the second signal curve being a representation of the concentration of the analyte of the standard sample detected by the chromatographic detector; measuring the first peak width W h i at the first absolute peak height h the second peak width J3 ⁇ 4 at the second absolute peak height h 2 , the third peak width J3 ⁇ 4 at the third absolute peak height h 3 , and the fourth peak width WM at the fourth absolute peak height h 4 of a peak in the second signal curve, wherein the first absolute peak height hi, the second absolute peak height h 2 , the third absolute peak height h 3 , and the fourth absolute peak height h 4 are different; and determining a peak shape index ratio of the peak in the second signal curve with the formula.
  • the method further comprises repeating the steps above on multiple injections of the standard sample; calculating a confidence range of the peak shape index ratio at a confidence level above 90% for the standard sample; and identifying the presence of the impurity in the sample where the determined peak shape index ratio of the sample is outside of the calculated confidence range.
  • the peak of the standard sample and the analyte peak of the sample under test have a same maximum peak height.
  • the method further comprises a suppressor coupled with the chromatographic column for receiving an output from the chromatographic column, wherein the suppressor is coupled with the chromatographic detector, such that an output from the suppressor is detected by the chromatographic detector.
  • a method of chromatographic quantitation of an analyte comprises flowing a first concentration of the analyte into a chromatographic column; detecting the analyte coming out from the chromatographic column by using a chromatographic detector; obtaining a first signal curve from the chromatographic detector, the first signal curve being a representation of the first concentration of the analyte detected by the chromatographic detector; determining a first width of a first peak in the first signal curve at a first absolute height of the first peak using a computing device; and quantifying the first concentration of the analyte based on the first determined width of the first peak.
  • the method further comprises setting the first absolute height to a value between 8 to 12 times a baseline noise level. In other embodiments, the first absolute height is approximately 60% of a maximum height of the first peak of the analyte. In some other embodiments, the method further comprises flowing the analyte at a second concentration into the chromatographic column; detecting the analyte coming out from the chromatographic column by using the chromatographic detector; obtaining a second signal curve from the chromatographic detector, in which the second signal curve also being a representation of the second concentration of the analyte detected by the chromatographic detector; determining a first maximum height of the first peak of the analyte in the first signal curve and a second maximum height of the second peak of the analyte in the second signal curve using the computing device; and setting the first, the second, or both absolute heights of the analyte to a value greater an 8 times a baseline noise level and less than a smallest of the first or second maximum
  • the first absolute height for the first determined width is the smaller of 55%-65% of the height of a peak maximum for the first peak and 55%-65% of the height of a peak maximum for the second peak.
  • the first signal curve represents a non-Gaussian peak.
  • the non-Gaussian peak is modeled by two separate Generalized
  • GSD Gaussian distribution
  • the determining the first width of the first peak comprises using independent exponential functions representing leading and trailing edges in the signal curve to model a peak.
  • the determining the first width of the peak is performed below a peak height accommodated by the first signal curve of the lowest analyte concentration of interest.
  • the determining the first width of the peak is performed at a peak height 60%-90% of a first maximum height of the peak of a lowest analyte concentration.
  • the first peak is clipped.
  • the method further comprises a suppressor coupled with the chromatographic column for receiving an output from the chromatographic column, wherein the suppressor is coupled with the chromatographic detector, such that an output from the suppressor is detected by the chromatographic detector.
  • a method of chromatographic quantitation of an analyte comprises flowing the analyte into a chromatographic column; detecting the analyte coming out from the chromatographic column by using a chromatographic detector; obtaining a signal curve from the chromatographic detector, the signal curve with a peak being a representation of the analyte detected by the chromatographic detector; fitting a height of the peak of the signal curve to an equation, the equation comprising:
  • a top equation describing a left half of the peak applies only at t ⁇ 0 while a bottom equation, describing a right half of the peak applies only at t > 0
  • h is the height of the peak
  • a maximum height of the peak appears at the intersection point of the above two equations
  • h max ,i is a maximum point in the top equation
  • h maXt2 is the maximum point of the bottom equation
  • m, n, a, and b are constants
  • the constants m, n, a and b are used to define a shape criterion for the peak.
  • the shape criterion is used for the identification of a peak.
  • the method further comprises determining a purity of the peak by taking 5% to 95% of the peak maximum to fit the pair of equations above.
  • the method further comprises determining an amount of impurity by deducting a maximum area that is fitted by using the pair of equations above from an area of the peak of the analyte detected.
  • the peak is quantitated on the basis of either of the two separate Gaussian distribution (GGD) functions, such that the concentration of the analyte is related by either a left half-width W h, i or a right half-width W h,r of the peak at any absolute height h;
  • GGD Gaussian distribution
  • the method further comprises a suppressor coupled with the chromatographic column for receiving an output from the chromatographic column, wherein the suppressor is coupled with the chromatographic detector, such that an output from the suppressor is detected by the chromatographic detector.
  • a system for chromatographic peak quantitation comprises a chromatographic column; a chromatographic detector configured to detect an amount of analyte from the chromatographic column; a signal converter converting the amount of an analyte detected to a signal curve; and an algorithm implemented computing device configured to determine a width of a peak in the signal curve in at least one selected height of the peak and quantify the amount of the analyte.
  • determining the width of a peak comprises determining the width of the peak in the signal curve at multiple heights of the peak.
  • the system further comprises a suppressor coupled with the chromatographic column for receiving an output from the chromatographic column, wherein the suppressor is coupled with the chromatographic detector, such that an output from the suppressor is detected by the chromatographic detector.
  • Figure 1A illustrates chromatographic system in accordance with some embodiments
  • Figure IB illustrates a flow chart of a width based single signal curve analyte quantitation method in accordance with some embodiments
  • Figure 1C illustrates a flow chart of a plurality signal curves determining (peak trace analysis) method 300 in accordance with some embodiments
  • Figure ID illustrates a flow chart of an impurity detecting method in accordance with some embodiments
  • Figure IE illustrates a peak trace analyzing method in accordance with some embodiments
  • Figure IF illustrates a peak trace analyzing method in accordance with some embodiments
  • Figure 1G illustrates a width-based analyte peak quantitation method in accordance with some embodiments
  • Figure 2 illustrates a plot of an error function in accordance with some embodiments
  • Figure 3 illustrates a plot of a Non-Gaussian peak generated by different functions in accordance with some embodiments
  • Figures 4A-4G illustrate some real chromatographic peaks of separated chemical components as well as fits computed by functions disclosed herein in accordance with some embodiments
  • Figure 5 illustrates a plot of relative bias and relative precision computed for a case of absorbance detection in accordance with some embodiments
  • Figure 6 illustrates a plot of relative error due to linear interpolation in accordance with some embodiments
  • Figure 7 illustrates a plot of relative error and relative standard deviation computed for width-based quantitation in accordance with some embodiments
  • Figure 8 illustrates the plot of Figure 7 in a magnified form in accordance with some embodiments
  • Figure 9A illustrates a plot of the sensitivity of a width measurement over a selected range in accordance with some embodiments
  • Figure 9B illustrates a logarithmic plot of the width measurement sensitivity data of Figure 9A in accordance with some embodiments
  • Figure 10 illustrates a plot of relative error and relative standard deviation computed for width-based quantitation in accordance with some embodiments
  • Figure 11 illustrates peak signal curve responses of certain chemical components produced in accordance with some embodiments
  • Figure 12 illustrates peak signal curve responses of certain chemical components produced in accordance with some embodiments
  • Figure 13 illustrates a nitrate chromatographic peak in accordance with some embodiments
  • Figure 14 illustrates a plot of a system responding nonlinearly at two different concentrations in accordance with some embodiments
  • Figure 15 illustrates a plot of conductometric responses in accordance with some embodiments
  • Figure 16 illustrates a Gaussian plot in accordance with some embodiments
  • Figure 17 illustrates a plot of leading and trailing half-widths for certain chemical components in accordance with some embodiments
  • Figure 18 illustrates plots for both the leading and trailing halves for certain analyte peaks in accordance with some embodiments
  • Figure 19 illustrates a plot of an analyte and an impurity peak in accordance with some embodiments
  • Figure 20 illustrates a plot of width vs. height for the situation of Figure 19 in accordance with some embodiments
  • Figure 21 illustrates a plot of an analyte and an impurity peak in accordance with some embodiments
  • Figure 22 illustrates a plot of a set of chromatograms for a bromide ion in accordance with some embodiments
  • Figure 23 illustrates a height-based calibration plot for the data in Figure 22 in accordance with some embodiments
  • Figure 24 illustrates a plot of peak shape conformity in accordance with some embodiments
  • Figure 25 illustrates a plot of peak shape conformity with offset corrections according to some embodiments
  • Figure 26 illustrates a plot of a set of chromatograms for bromide samples in accordance with some embodiments
  • Figure 27 illustrates a plot of the data for chloride of Figure 11 in accordance with some embodiments
  • Figure 28 illustrates a plot of the intercepts of the data of Figure 27 in accordance with some embodiments
  • Figure 29 illustrates a plot of chromatographic data for caffeine in accordance with some embodiments
  • Figure 30 illustrates a plot of linear correspondence in accordance with some embodiments
  • Figure 31 illustrates a plot of chromatographic data impurity detection in accordance with some embodiments
  • Figures 32A and 32B illustrate paired plots of the separation of isomers by Gas Chromatography Vacuum Ultraviolet Spectroscopy on the left panel and purity analysis plots for the same on the right in accordance with some embodiments;
  • Figure 33 illustrates a plot of normalized spectra obtained from peak height maxima at different wavelengths in accordance with some embodiments
  • Figure 34 illustrates a plot of spectrum reconstruction in accordance with some embodiments
  • Figure 35 illustrates a plot of the same data as Figure 34 with multipliers applied in accordance with some embodiments
  • width based quantitation can provide a near-infinite number of calibration equations. Spectrum reconstruction of a truncated peak due to detector saturation is possible through width considerations. While this can also be done by other means, the width based approach may readily provide clues to the presence of an impurity.
  • Embodiments of this disclosure entail WBQ techniques.
  • WBQ can offer superior overall performance (lower root mean square error over the entire calibration range compared to area or height based linear regression method), rivaling 1/x 2 - weighted linear regression.
  • a WBQ quantitation model is presented based on modeling a chromatographic peak as two different independent exponential functions which respectively represent the leading and trailing halves of the peak. Unlike previous models that use a single function for the entire peak, the disclosed approach not only allows excellent fits to actual chromatographic peaks, it makes possible simple and explicit expressions for the width of a peak at any height. WBQ is applicable to many situations where height or area based quantitation is simply inapplicable.
  • the disclosed WBQ embodiments present a general model that provides good fits to both Gaussian and non-Gaussian peaks without having to provide for additional dispersion and allows ready formulation of the width at any height.
  • peak width is measured at some fixed height (not at some fixed fraction of the peak maximum, such as asymmetry that is often measured at 5% or 10% of the peak maximum).
  • This disclosure relates generally to methods of analyzing data obtained from instrumental analysis techniques used in analytical chemistry and, in particular, to methods (and related systems and devices) of automatically identifying peaks in liquid chromatograms, gas chromatograms, mass chromatograms, flow-injection analysis results (fiagrams), electropherograms, image-processed thin-layer chromatograms, or optical or other spectra.
  • methods and related systems and devices of automatically identifying peaks in liquid chromatograms, gas chromatograms, mass chromatograms, flow-injection analysis results (fiagrams), electropherograms, image-processed thin-layer chromatograms, or optical or other spectra.
  • Figure 1A depicts a chromatographic system 100 in accordance with some embodiments.
  • the system 100 comprises a controlling and computing device 102, a detecting unit 104, a suppressor unit 106, a separation unit 108 (e.g., chromatographic column), a delivery unit 110 (e.g., pump), and a solvent providing unit 112 (e.g., an eluent providing system or container).
  • a controlling and computing device 102 the system 100 comprises a controlling and computing device 102, a detecting unit 104, a suppressor unit 106, a separation unit 108 (e.g., chromatographic column), a delivery unit 110 (e.g., pump), and a solvent providing unit 112 (e.g., an eluent providing system or container).
  • a separation unit 108 e.g., chromatographic column
  • delivery unit 110 e.g., pump
  • a solvent providing unit 112 e.g., an eluent providing system or
  • the controlling and computing device 102 contains a processor and memory.
  • the device 102 is implemented with executable computing instructions for performing a predetermined specific functions.
  • the executable computing instructions are compiled or structured as a computer software, which configures the processor and the electron storing structures to store and locate voltages for performing a predetermined functions according to the loaded algorithm (e.g., the peak width determining algorithm disclosed herein).
  • the controlling and computing device 102 control s/commands the performance of the system 100.
  • the detecting unit 104 comprises a chromatography detector, including destructive and non-destructive detectors.
  • the destructive detectors comprise a charged aerosol detector (CAD), a flame ionization detector (FID), an aerosol-based detector (NQA), a flame photometric detector (FPD), an atomic- emission detector (AED), a nitrogen phosphorus detector (PD), an evaporative light scattering detector (ELSD), a mass spectrometer (MS), an electrolytic conductivity detector (ELCD), a sumon detector (SMSD), a Mira detector (MD).
  • CAD charged aerosol detector
  • FID flame ionization detector
  • NQA aerosol-based detector
  • FPD flame photometric detector
  • AED atomic- emission detector
  • PD nitrogen phosphorus detector
  • ELSD evaporative light scattering detector
  • MS mass spectrometer
  • ELCD electrolytic conductivity detector
  • SMSD sumon detector
  • MD Mira detector
  • the nondestructive detectors comprise UV detectors, fixed or variable wavelength, which includes diode array detector (DAD or PDA), a thermal conductivity detector (TCD), a fluorescence detector, an electron capture detector (ECD), a conductivity monitor, a photoionization detector (PID), a refractive index detector (RI or RID), a radio flow detector, a chiral detector continuously measures the optical angle of rotation of the effluent.
  • the separation unit 108 comprises a chromatographic column.
  • the chromatographic column is able to be liquid chromatographic column, gas chromatographic column, and ion-exchange chromatographic column. A person of ordinary skill in the art will appreciate that any other chromatographic column is within the scope of the present disclosure, so long as the chromatographic column is able to be used to separate one analyte from another.
  • Figure IB illustrates a flow chart of a width based single signal curve analyte quantitation method 200 in accordance with some embodiments.
  • a sample is prepared and injected into a chromatography (e.g., ion-exchange chromatography) with a predetermined condition (e.g., 65°C at a flow rate of 0.5 mL/min.)
  • a signal curve is obtained using a chromatographic detector, the curve being a representation of at least one analyte component detected by the detector.
  • a mathematical computation is performed using the signal curve via the computing device described above with one or more implemented algorithms disclosed herein, wherein the computation comprises determining the width of a peak in the curve in at least one selected height of the peak.
  • the determined width is used to determine a characteristic associated with the at least one analyte component.
  • FIG. 1C illustrates a flow chart of a plurality of signal curves determining (peak trace analysis) method 300 in accordance with some embodiments.
  • a sample is prepared and injected into a chromatography with a predetermined condition.
  • a plurality of signal curves are obtained using a detector, each curve being a representation of a concentration of at least one analyte component detected by the detector.
  • a mathematical computation is performed using the plurality of signal curves, wherein the computation comprises determining the width of a peak in each curve at a selected height of the respective peak.
  • the determined peak widths are used to produce at least one calibration curve.
  • Figure ID illustrates a flow chart of an impurity detecting method 400 in accordance with some embodiments.
  • a sample is prepared and injected into a chromatography with a predetermined condition.
  • a signal curve is obtained using a chromatographic detector, the curve being a representation of a chemical mixture detected by the detector.
  • a mathematical computation is performed using the signal curve, wherein the computation comprises determining the respective width of a peak in the curve at a plurality of selected heights of the peak.
  • the determined widths are used to detect an impurity in the mixture.
  • Figure IE illustrates a peak trace analyzing method 500 in accordance with some embodiments.
  • a sample is prepared and injected into a chromatography with a predetermined condition.
  • a signal curve is obtained using a detector.
  • a fitting of a peak in the signal curve is determined by performing a mathematical computation, wherein the computation comprises independently fitting each side of the peak with a generalized Gaussian distribution function.
  • Figure IF illustrates a peak trace analyzing method 600 in accordance with some embodiments.
  • a sample is prepared and injected into a chromatography with a predetermined condition.
  • a signal curve is obtained using a detector, the curve being a representation of at least one analyte component detected by the detector.
  • a mathematical computation is performed using the signal curve, wherein the computation comprises determining the widths of a peak in the curve at a plurality of selected heights of the peak.
  • the determined peak widths are used to determine a shape criterion for the peak. The computations are performed via the techniques disclosed herein.
  • FIG. 1G illustrates a width -based analyte peak quantitation method 700 in accordance with some embodiments.
  • a Step 702 it is determined if the peak maximum reaches a nonlinear or saturated detector response region.
  • the process goes to Step 704 if it is determined that the peak maximum reaches a nonlinear or saturated detector response region, and the process goes to Step 706 if it is determined that the peak maximum does not reach a nonlinear or saturated detector response region.
  • the width is measured at a signal height, where the detector is not saturated or nonlinear. In some embodiments, the height is chosen as high as possible in the permissible range.
  • the height is chosen at a height where calibration has already been computed. In some other embodiments, the height is chosen where the width of the unknown peak is measured. Next, a calibration curve is constructed at that height from stored calibration peaks. Next, the calibration curve is used to interpret the concentration of the unknown. At the Step 706, the width of the peak is measured at the greater of 60% of the peak maximum or at a height 20x the baseline noise level but not exceeding 95% of the peak maximum. Next, a calibration curve is constructed at that height from stored calibration peaks. Next, the calibration curve is used to interpret the concentration of the unknown. Alternatively, a height is chosen for quantitation for which a calibration already exists as long as it is not below 5% of the peak height or 20x the baseline noise level.
  • the disclosed Width-Based Quantitation (hereinafter "WBQ") measuring methods and devices are applicable to both Gaussian and non-Gaussian peaks of one or more analytes from a chromatography device, with the merit that the resulting RMS errors are comparable to those using height or area-based quantitation using weighted regression.
  • Embodiments of the disclosed WBQ method, process, and system may also be used as a complement to conventional techniques: quantitation can be height-based at the low-end, width-based at the high end (where detector saturation/nonlinearity may set in) and area-based at intermediate concentrations.
  • WBQ provides notable advantages, including: (a) lower overall RMS error without weighting compared to unweighted area or height based quantitation, (b) applicability over a large range of concentrations, (c) accurate quantitation when (i) the detector response is in the nonlinear response range, (ii) the detector response is saturated at the high end, and (iii) the detector response is not a single valued function of concentration, and (d) detection of co- eluting impurities, none of which situations can be handled by area or height-based quantitation.
  • the height h at which width is being measured is low enough to be in the linear response domain of the detector/analyte/column system.
  • the ascending peak has no foreknowledge of whether the peak maximum will remain within the linear response domain, or in the extreme case, become completely clipped.
  • h max computed from Equation (4) is the height that would have been registered if the analyte peak remained within the linear domain, regardless of whether it actually was or not.
  • h max is therefore linearly related to the concentration C, providing a more general form of Equation (4):
  • Non-Gaussian Peaks have been modeled as exponentially modified Gaussian (EMG) or polynomial modified Gaussian (PMG) peaks. The width at a particular height for a specific EMG function is easily numerically computed.
  • EMG exponentially modified Gaussian
  • PMG polynomial modified Gaussian
  • top equation pertains to one half of the peak and the bottom to the other:
  • Equations (7) - (9) There are limitations on the ranges of parameters in Equations (7) - (9) that can be easily imposed.
  • a consideration of peak shapes of the exponential functions in Equations (7) - (8) will indicate that for real chromatographic peaks the values of m and n would usually lie between 1 and 2, the reciprocals 1/m and 1/n therefore lie between 1 and 0.5.
  • Equation (9) can be readily expressed reciprocally as which has obvious bounds of 0 and 1, more typically between 0.05 and 0.95, meaning width is to be measured between the bounds of 5% and 95% of h max with the only modification of a negative sign before the logarithmic terms. With these constraints, it is readily shown numerically (see the following mathematical calculations) that the sum expression in Equation (9) above can always be expressed by a single similar term as in Equation (10) below, with ⁇ 1% root-mean-square error (RMSE), at least within the domain of 1.05-20 (h being 5-95% of peak maximum).
  • RMSE root-mean-square error
  • Figure 2 illustrates a plot of L2 error function in the parameter space (c, r) for the region [0, l]x[0,l] in accordance with some embodiments.
  • the error function in the region ⁇ (c, r) G [0, 1] X [0, 1] ⁇ , we compute the error function as depicted in Figure 2.
  • Figure 4A illustrates the fit of the 1 mM chloride to Equation (9) analog.
  • Figure 4B illustrates the fit of the 6 mM nitrate to Equation (9) analog.
  • the chromatographic conditions for the chloride fit are illustrated in Figure 11 and in Figure 12 for the nitrate fit. From 1 %-99% of peak height, RMSE as a percentage of h max : Chloride: 0.66% (r 2 0.9996), Nitrate: 1.2% (r 2 0.9987).
  • Figure 4C illustrates the fit of experimental 5 mM citrate peak to Equation (9) analog.
  • the chromatographic conditions being as illustrated in Figure 11, the RMSE as a percentage of h max : 0.55% (r 2 0.9998).
  • Figures 4D and 4E illustrate the Equation (9) analog fits for 6 mM formate.
  • Figure 4D has the best fit using the data for the entire peak. This fit is obviously poorer at the low and especially high h extremes compared to Figures 4A-4C. Considering that neither extreme of height will typically be used for WBQ, it makes more sense to fit the curve excluding the extremes, e.g., as in Figure 4E, where only the time intervals that comprises 5- 95% of the peak height in the original data are used.
  • the RMSE as a fraction of h max improves from 2.4 to 1.4%; r 2 improves from 0.9944 to 0.9975.
  • Figures 4F-4G illustrate the Equation (9) analog fits for 2 mM acetate. Similar to Figures 4D-4E, Figure 4F has the best fit using the data for the entire peak. In Figure 4G, only the time intervals that comprise 5-95% of the peak height in the original data are used.
  • the RMSE as a fraction of h max improves from 1.2 to 0.81%; r 2 improves from 0.9986 to 0.9991.
  • h max can in this case be then expressed as:
  • the value of n' is equal to n. In other embodiments, the value of n' is different from n. In some embodiments, n' is a constant like n.
  • Embodiments of this disclosure entail the detection of the beginning and the end of a peak, generally through the specifications of a threshold slope or a minimum area of a peak. Finding the height maximum is thereafter straightforward as it corresponds to the maximum value observed within the domain of the peak so-defined. However, the measured maximum is affected by the noise and that translates both into inaccuracy and uncertainty. To simulate random noise, the results below represent 10,000 trials.
  • the error in the average height (consider this as the bias or accuracy) ranges from -1.7% at 10 Hz to +1.6% at 50 Hz, the errors are a combined result of inadequacy of sampling frequency (this is the dominant factor at low sampling rates), noise and stray light; the relative SD ("RSD") of this perceived height (the uncertainty) is quite low and is in the 0.3-0.4%) range from 10-50 Hz.
  • Figure 5 illustrates the relative bias (solid lines, left ordinate) and relative precision (dashed lines right ordinate) computed for a case of absorbance detection in accordance with some embodiments.
  • the situation assumes a Gaussian analyte peak with a true absorbance amplitude of 1 mAU, a SD of 1 s, 20 ⁇ of peak to peak random noise atlO Hz and 0.05%> stray light.
  • the results shown depict averages and SDs of 10,000 computational trials.
  • 502, 504 and 506 traces resectively depict height, width, and area-based quantitation; width measured at 150 ⁇ . Both bias and precision improves as absorbance increases until bias is affected by the stray light.
  • Some embodiments to determine the width at a given height first proceed to determine the location of the specified height h on the signal curve on the ascending and descending edges of the signal and determine the times ti and t 2 corresponding to h, and hence determine W h as t 2 -ti.
  • Figure 6 depicts the relative error due to linear interpolation as a function of 1/ ⁇ , assuming no noise in accordance with some embodiments.
  • the black or red error curves touch the blue zero error line, the width is being measured across points actually sampled where no interpolations are needed.
  • the error decreases, with the minimum error being reached at an abscissa value of -0.6; the direction of the error changes thereafter.
  • additional errors arise, first in locating h.
  • Figure 5 illustrates the relative error in h max computed based on the width-based quantitation using Equation (5) for the same base case as above as a function of / and ranges from -1.4% at 10 Hz to ⁇ 0.3% at 50 Hz, better than that based strictly on height ( Figure 5). But at 2-3%) RSD, uncertainties in this range are significantly higher than either height or area based quantitation, although hardly in the unacceptable range considering the width measurement is actually being made at a height below the limit of quantitation (LOQ, at 10 times the noise level this would be 200 ⁇ ).
  • the bias and precision are already -0.5 % and 0.7%, respectively at a sampling frequency of 20 Hz (See Figure 7)
  • Figure 7 illustrates the relative error (solid lines, left ordinate) and RSD (dashed lines right ordinate, note logarithmic scaling) computed for a case of absorbance detection and WBQ.
  • the situation assumes a Gaussian analyte peak with a true absorbance amplitude of 1, 10, 100, 1000, and 10,000 mAU (red 702, blue 704, green 706, purple 708, and orange 710 traces respectively), all measured at 1/h of 0.15, a SD of 1 s.
  • the peak to peak random noise is 20 ⁇ atlO Hz and corresponding noise values under other conditions.
  • the stray light is assumed to be 0.05%.
  • the results shown depict averages and SDs of 10,000 computational trials.
  • the black trace indicates the 1 mAU case without any noise. While the 1 mAU case without noise displays an RSD, it has an RSD higher than all the other higher absorbance traces that do include noise. This is because the interpolation errors are still present and are relatively much greater at lower absorbances. The relative errors are also illustrated in Figure 8 in a magnified form over a more limited range of
  • Figures 9A-9B illustrate the sensitivity of the width measurement due to uncertainty in height in two different ways in accordance with some embodiments.
  • Figure 9A covers the primary range of interest, 5% to 95% of peak height; the negative sign of the ordinate values results from the fact that width always decreases with increasing height, the absolute values have been multiplied by 100 to indicate percentage dependence. The magnitude of this sensitivity increases steeply at either end.
  • Figure 9B illustrates a plot of the log of dW h /dh after changing its sign (to permit logarithmic depiction) vs. 1/h.
  • Figure 10 illustrates the relative error (or relative bias, solid lines, left ordinate) and RSD (or relative precision, dashed lines right ordinate) computed for a case of absorbance detection and WBQ in accordance with some embodiments.
  • the situation assumes a Gaussian analyte peak with a true absorbance amplitude of 1 mAU, a SD of 1 s, 20 ⁇ of peak to peak random noise at 10 Hz and 0.05% stray light.
  • the results shown depict averages and SDs of 10,000 computational trials. Red 1002, purple 1004 and brown 1006 traces respectively measured at of 0.15, 0.60 and 0.85.
  • the setup entailed a ThermoFisher/Dionex: IC-25 isocratic pump, EG40 electrodialytic eluent generator, 2 mm bore AG20/AS20 guard and separation column, LC30 temperature controlled oven (30 °C), ASRS-Ultra II anion suppressor in external water mode, CD-25 conductivity detector.
  • An electrogenerated KOH gradient at 0.25 mL/min was used as follows: Time, min (Concentration, mM): 0(4), 3(4), 15(10), 19(40), 27(40), 27.5(4), 30(4).
  • formate, trifluoroacetate and nitrate eluted under a specific gradient condition show extensive tailing and/or fronting.
  • the experimental setup relating to Figure 12 was similar to that of Figure 11, except for KOH eluent: (0.3 mL/min) 0-10 min, 2.0 mM; 10-15 min, 2.0-10 mM; 15-32, min, 10 mM.
  • Table 1A Weighted and Unweighted %RMS Errors. Area, Height, Width based Quantitation. (Near)-Gaussian Peaks
  • Tailing/Fronting Peaks Because of variable dissociation of weak acid analytes and the interplay of both electrostatic and hydrophobic retention mechanisms where gradient elution largely alters only the electrostatic push, non-Gaussian peaks are common in ion chromatography (IC) ( Figure 11). Width was measured at 3.0, 1.5, and 2.0 ⁇ 8/ ⁇ for formate, trifluoroacetate and nitrate, respectively, substantially above the baseline noise levels but still below the height of the lowest concentration peak in each case.
  • WBQ substantially outperforms area or height based quantitation by unweighted regression and rivals l/x 2 -weighted regression.
  • WBQ can also benefit post-column reaction based detection methods which exhibit a finite detector background from the post- column reagent because it is not necessary to have a stoichiometric amount of the post- column reagent to accommodate the highest analyte concentration of interest.
  • WBQ can make use of the two-dimensional nature of chromatographic data: If multiple heights are used for quantitation or if used in conjunction with height or area based quantitation it is possible to check for and detect co-eluting impurities.
  • Nonlinear response situations include scenarios where the detector response is not a single valued response of concentration, a notable example being fluorescence behavior of a fluor at high enough concentrations in the self-quenched domain. While such phenomena have occasionally been used advantageously in indirect fluorometric detection using fluorescent eluents at high concentrations to produce positive signals, a fluorescent substance with a peak concentration in the self-quenched domain will produce an M-shaped peak.
  • a single quantitation paradigm involving both the low concentration unquenched and the higher concentration self-quenched domain has not been possible. Similar situations may be encountered in post-column reaction detection. WBQ can be applied in these situations to provide accurate quantitation.
  • Width can be measured at many heights.
  • Co-eluting impurities by definition are smaller than the principal component in the peak, and therefore contribute to a greater degree to the peak width towards the bottom than towards the top. As such, the presence of an impurity may not be readily apparent from asymmetry changes. But, if the concentration of the examined band is ascertained by a calibration curve generated from pure standards, the telltale indication of an impurity is a significantly higher predicted concentration when interpreted with a width-based measurement at a lower height compared to one at a higher height.
  • the width of the left half and the right half can be independently measured and their depiction as a function of height directly (or in a transformed form) provides information about asymmetry and other characteristics of the band not available from any single parameter description of peak asymmetry.
  • Figure 13 illustrates a nitrate peak chromatogram detected at 200 nm with the concentration spanning two orders of magnitude on an Agilent 1290 DAD instrument - chromatographic details ICS 5000 IC system: AG11 (2 x 50 mm)+ ASH (2x 250 mm) columns. KOH gradient at 0.3 mL/min: 0-10 min, 2.0 mM; 10-15 min, 2.0-10 mM; 15-32 min, 10 mM; Injection volume, 10 ⁇ The width was measured at 20 mAU, far above baseline noise, and could be represented by the equation (in the form of Equation (16)):
  • PCR post-column reagent
  • a unique relevant example is the detection of acidic eluites by introducing a small amount of a base post- column (the column background is pure water) and then allowing the mixture to flow through a conductivity detector, which we have explored for some time.
  • the detector background reflects the conductivity from the base added; when an acid eluite comes out, the acid HX is neutralized forming X " and water. The net result is thus the replacement of OH " by X " .
  • OH " has the highest mobility of all anions, a negative response in the conductivity baseline results. However, if the eluite acid concentration exceeds the base concentration, the conductivity will go back up as the peak concentration is approached.
  • Figure 14 illustrates two different concentrations of H 2 S0 4 injected into a 100 ⁇ strong base carrier in accordnace with some embodiments.
  • a negative peak results (red trace 1402).
  • a W-shaped peak results (black trace 1404).
  • a fluorescent substance is injected into a nonfluorescent carrier, and the resulting signal monitored with a fluorescence detector, an M-shaped peak will result if the fluorescence is in the self-quenched domain at the peak. Both belong to a general case where the response is not a single-valued function of the concentration.
  • Figure 15 illustrates conductometric responses of two anions, each over two orders of magnitude, to a detection system using a permeative amine introduction system (PAID) in accordance with some embodiments.
  • Formic acid is moderately weak (pK a 3.75); trifluoroacetic acid is almost a strong acid (pK a 0.25). These responses cannot be quantitated by height or area-based methods.
  • the depicted set of illustrative data is from a post-column reaction system where a base (Et 2 NH 2 OH) is introduced to react with formic and trifluoroacetic acid eluites to produce the resulting salt that is detected in a background of base.
  • Et 2 NH 2 OH itself has a measurable detector background, it is desirable to minimize the added amount to reduce the baseline noise.
  • the width was measured at a fixed height (1.2 and 0.4 ⁇ 8/ ⁇ for formate and trifluoroacetate, respectively) below the baseline to construct a calibration plot.
  • the Relative RMSE over the two orders of magnitude range of concentration was 6.6% for formate and 13.6% for trifluoroacetate.
  • the error is relatively high in the second case because the peak shape actually changes at the higher concentrations.
  • WBQ still provides a viable option.
  • the orange solid trace and the dashed black trace in the main plot respectively shows the left and right half width for this peak as a function of l/h (h being h max /h).
  • the two halves are mirror images and the half-width plots therefore appear superimposed.
  • SD standard deviation
  • the circles representing the leading half of the top peak completely overlap the previous half-width vs. l/h traces.
  • the right half of the top peak provides a very different half-width vs. l/h trace. This figure emphasizes that peak symmetry (or lack thereof) is much easier to ascertain in combined left half-width and right half-width vs. l/h plots than in the original chromatographic peaks.
  • WBQ is essentially a depiction of width as a function of height.
  • a clear visual depiction of asymmetry appears if the left and right half widths are independently shown as a function of height.
  • the l/h bounds are deliberately limited to 0.05 to 0.95 (in our experience, conformity of each side of real chromatographic peaks to a generalized Gaussian distribution model (GGDM) is better attained within these limits (See Figures 4D-4G)). As such, in these depictions, greater concern was given to the general shape of the peak rather than fronting or tailing only near the peak base.
  • GGDM generalized Gaussian distribution model
  • Figure 17 indicates how this type of depiction reveals symmetry in real peaks in accordance with some embodiments.
  • Figure 17 illustrates the left (leading) and right (trailing) half-width vs. l/h plots for acetate, formate, chloride, nitrite, nitrate, and citrate.
  • the original chromatograms can be seen in Figures 4A-4G. Note that while generally the trailing half is wider than the leading half, it is the reverse for the formate peak which fronts quite obviously.
  • the absolute value of the width is dependent on the SD of the peak and the injected concentration. With the exception of formate, which has a strongly fronting peak, the trailing halves are always wider than the leading halves.
  • Figure 15 depicts, for a Gaussian peak, a plot of W h vs. l/h is not expected to be linear, but departs increasingly from linearity as l/h decreases.
  • a linear W h vs. l/h plot connotes a triangle, a near-triangular shape can be seen for the strongly fronting leading edge of the formate peak).
  • Nitrate exhibits the largest asymmetry; the trailing half is much wider than the leading half throughout and increasingly so with decreasing l/h.
  • the exponent of In h can have a value m other than 0.5.
  • the departure from the ideal Gaussian distribution can be judged from how far m departs from 0.5 (illustrative distributions are illustrated in Figures 1A-1B, exponent n in these figures equals llm).
  • a logarithmic transformation of Equation (18) produces a linear form: A plot of In W h( i, t) as a function of In (In h) thus produces m as the slope and the SD s is given by 0.707*exp(intercept).
  • Figure 18 illustrates the relevant plots for both the leading and trailing halves for the five analyte peaks (shown in Figures 4A-4G) along with the slope (an index of departure from true Gaussian distribution) and the coefficient of determination (an index of conformity to GGDM), and the SD (an index of the width of the corresponding halves of the peak).
  • the circles represent the leading edge
  • dashed lines represent the trailing edge.
  • No conventional method can detect low levels of an impurity that has similar detection characteristics (e.g., identical absorption spectrum) that appear at an identical retention time as the analyte.
  • WBQ is capable of detecting a difference in principle even under these conditions if there is a difference in peak shape between the impurity and the analyte.
  • the sum of (a) and (b) thus results in (c).
  • the respective SDs are 1.41 and 0.71 units.
  • dashed blue (f) and dashed orange traces (g) are respectively the sum of (d) and (e), with
  • WBQ predictions for a lower SD impurity at a higher l/h will mean a greater relative change in concentration compared to that at a lower l/h and the reverse would be the case when the SD of the impurity is higher than that of the analyte.
  • increasing SD of the impurity will increase the width and thence the concentration prediction more and more at lower l/h values, whereas near the apex the contribution of the impurity will remain the same if its amplitude remains the same.
  • Figure 20 illustrates a width vs. height plot for the situation in Figure 19 in accordance with some embodiments; the same conclusions are reached.
  • the corresponding value for the suspect peak can then be compared with that for the standard(s), including the uncertainty and it can be determined whether within the desired limits of uncertainty the suspect peak falls within the expected shape parameter range. As this approach does not require detailed calibration curves, its use is illustrated below in impurity detection.
  • FIG. 22 illustrates a set of chromatograms from an injection of bromide ion at concentrations of 200, 500, 800, 1000 and 2000 ⁇ in a typical suppressed ion chromatographic setup in accordance with some embodiments.
  • Chromatographic conditions ThermoFisher Dionex ICS-5000 system, AG20 (2 x 50 mm) + AS20 (2 x 250 mm), Electrogenerated KOH eluent 8.0 mM, 0.25 mL/min, Dionex AERS 500 2 mm suppressor.
  • Figure 26 illustrates the chromatograms for the same bromide samples except that they now contain the same constant concentration of nitrate (20 ⁇ ) as impurity, the relative amount thus being 1-10% on a relative molar basis. With the possible exception of the 10% case, the presence of the impurity is not readily discernible by visual examination. Table 6 below presents data on impurity detection based on this criterion (the numbers in red in the last column indicate values outside the 95% confidence range and hence that peak shapes are different from that of the standards: an impurity may be present).
  • FIG. 27 illustrates the data for chloride in Figure 11 along with the best fit equations in accordance with some embodiments; the data is plotted in the form of Equation (20), a linear plot results throughout.
  • the intercept b in Equation (20) should be linearly proportional to In C provided the data is entirely in the linear response domain.
  • Figure 28 illustrates the plot of the intercepts (b) in Figure 27 vs. In C, displaying that the correspondence of b with In C also holds an excellent linear correlation.
  • Figure 30 illustrates that the linear correspondence between the intercept b and In C also breaks down in accordance with some embodiments. This plot is similar to that in Figure 28. When all the data (red triangles) are used in the regressions done in Figure 29, deviation from linearity at the higher concentration end is evident. If the data in the top left quadrant of Figure 29 are omitted before performing regression (black circles), a much better linear fit is obtained. In the caffeine chromatograms, the impurity is invisible when the caffeine peak is plotted to accommodate the maximum peak height but becomes readily apparent when observed in a magnified view of the baseline being measured.
  • Figure 31 illustrates how the trace at 20,000 ng (red trace A) would appear to be a perfectly normal peak.
  • the detector is predictably saturated at 10,000 ng (black trace B), but no abnormality is readily evident on its tail; only when examined at a high magnification (blue trace C), the appearance of an impurity peaking at just over 3.2 min becomes readily apparent. It is clear that width measurement at low heights in the present case will lead to error. Indeed, the failure of the caffeine data to fit the general model in Equations (10) and (21) - (22) is what led us to examine the baseline of the high concentration peaks in greater detail. In other words, efforts to WBQ quantitation led to the detection of this impurity, of which those that generated the data were unaware.
  • Figure 32A illustrates the separation of Dimethylnaphthalene (DMN) isomers by Gas Chromatography Vacuum Ultraviolet Spectroscopy (from Schenk, J.; Mao, J. X.; Smuts, J.; Walsh, P.; Kroll, P.; Schug, K. A. Anal. Chim. Acta 2016, 945, 1-8).
  • the mixture contained 10% 1,4-DMN and 90% 2,3-DMN.
  • the circles represent the detector response with the red and blue lines representing best estimates on the response of each isomer based on spectral deconvolution.
  • Figure 32B illustrates independent left and right edge In W h vs. In (In h) plots to aide in the detection of impurity.
  • a photodiode array UV-VIS absorbance detector is one of the most common detectors used in high performance liquid chromatography (HPLC) and has the capability of providing an absorption spectrum of the analyte "on the fly", by taking a spectral snapshot as the eluite passes through the detector. As the absorption spectrum is unique to a particular molecule, availability of the spectrum aids in eluite identification or confirmation of the putative identity.
  • a process for obtaining the spectrum is to simply plot the maximum absorbance (peak height) observed at different wavelengths as a function of the wavelength and this may be then optionally normalized by dividing by the sample volume (or mass, if known) injected.
  • Figure 33 illustrates normalized spectra obtained from peak height maxima at different wavelengths as described above when different amounts of solute (caffeine) are injected in a chromatographic system.
  • the spectra at 800 ng and 2000 ng are completely overlapped and do not show any evidence of detector saturation.
  • Peak maxima absorbance-based reconstruction can be done in other ways. One approach is to move away from the peak maximum to a location on the rising or the trailing edge of the peak where there is no saturation.
  • the depiction of a peak with time as the abscissa and l/h (rather than absolute absorbance) as the ordinate is identical for wavelengths in the absence of detector aberrations.
  • the absorbance maximum can be calculated from any value of t where the absorbance is within the normal range and the previously determined value of l/h at that value of t.
  • Equation (8) Spectral reconstruction based on shape recognition/WBQ embodiments can be carried out in several ways, all based on the implicit basis of WBQ that the GGDM fits one or both edges of the peak as given in Equation (8). If the chromatographic peak for the non- truncated peak is presented as l/h vs. time, the appropriate form of Equation (8) here will be:
  • the best fit of the h vs. t data to Equation (24) is sought by varying h max , which is implicit in h using any nonlinear fitting routine e.g., Microsoft Excel SolverTM.
  • h max implicit in the expression of h is h max , which is not known.
  • the fitting routine then varies h max simultaneously for both the ascending and descending sides of the data to obtain the best linear r 2 .
  • the methods and devices are used to separate a sample with one or more chemical substances and determine the concentration of each of the chemical substances using the width-based quantitation algorithm implemented computing device and methods.
  • an amount of analyte is detected by a detector after passing through a chromatography column, the amount of analyte detected is converted to a signal curve (e.g., a peak shape), and a width-based quantitation algorithm is used to determine a concentration of the analyte of the signal curve.
  • a signal curve e.g., a peak shape
  • any embodiment referenced herein is freely combinable with any one or more of the other embodiments referenced herein, and any number of features of different embodiments are combinable with one another, unless indicated otherwise or so dictated by the description herein.
  • This disclosure may include descriptions of various benefits and advantages that may be provided by various embodiments. One, some, all, or different benefits or advantages may be provided by different embodiments.
  • processors may be implemented as software constructs stored in a machine accessible storage medium, such as an optical disk, a hard disk drive, etc., and those constructs may take the form of applications, programs, subroutines, instructions, objects, methods, classes, or any other suitable form of control logic; such items may also be implemented as firmware or hardware, or as any combination of software, firmware and hardware, or any combination of any two of software, firmware and hardware.
  • a machine accessible storage medium such as an optical disk, a hard disk drive, etc.
  • processors may refer to one or more processors.
  • an article of manufacture comprises a non-transitory machine-accessible medium containing instructions, the instructions comprising a software application or software service, wherein the instructions, when executed by the machine, cause the machine to perform the respective method.
  • the machine may be, e.g., a processor, a processor-based system such as the systems described herein, or a processor-based device such as the user interface devices described herein.

Abstract

La présente invention concerne des systèmes, des procédés et des dispositifs pour fournir des procédés analytiques pour des réponses sous forme de pics séparés dans le temps ou l'espace, comprenant la quantification de pics chromatographiques sur la base de la mesure de la largeur d'un tracé de pic à une hauteur sélectionnée en tant qu'élément de quantification. L'invention concerne en outre des procédés de traitement d'un tracé de pic en tant que composition de fonctions exponentielles représentant une extrémité avant et une extrémité arrière. L'invention concerne en outre des procédés qui permettent la détection d'impuretés dans des sorties de tracé de pic.
PCT/US2017/063536 2016-11-28 2017-11-28 Systèmes, procédés et dispositifs pour l'analyse à base de largeur de tracés de pics WO2018098502A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201780084712.8A CN110234990B (zh) 2016-11-28 2017-11-28 用于基于宽度的峰迹线分析的系统、方法和装置
EP17874213.6A EP3545296A4 (fr) 2016-11-28 2017-11-28 Systèmes, procédés et dispositifs pour l'analyse à base de largeur de tracés de pics

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201662427119P 2016-11-28 2016-11-28
US62/427,119 2016-11-28

Publications (1)

Publication Number Publication Date
WO2018098502A1 true WO2018098502A1 (fr) 2018-05-31

Family

ID=62190755

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2017/063536 WO2018098502A1 (fr) 2016-11-28 2017-11-28 Systèmes, procédés et dispositifs pour l'analyse à base de largeur de tracés de pics

Country Status (4)

Country Link
US (1) US10634653B2 (fr)
EP (1) EP3545296A4 (fr)
CN (1) CN110234990B (fr)
WO (1) WO2018098502A1 (fr)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11467140B2 (en) * 2017-11-15 2022-10-11 Mecanique Analytique Inc. Method and apparatus for analysing a gas chromatography elution peak
US11244818B2 (en) * 2018-02-19 2022-02-08 Agilent Technologies, Inc. Method for finding species peaks in mass spectrometry
EP4110492A1 (fr) * 2020-02-28 2023-01-04 Waters Technologies Corporation Introduction d'échantillon réglable
US20220042957A1 (en) * 2020-08-04 2022-02-10 Dionex Corporation Peak Profile for Identifying an Analyte in a Chromatogram
US11860143B2 (en) * 2021-12-21 2024-01-02 Dionex Corporation Method and system for self-regulating a suppressor

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4468742A (en) * 1981-03-17 1984-08-28 The Regents Of University Of California Microprocessor system for quantitative chromatographic data analysis
US5121443A (en) * 1989-04-25 1992-06-09 Spectra-Physics, Inc. Neural net system for analyzing chromatographic peaks
US6438499B1 (en) 1998-09-10 2002-08-20 Tosoh Corporation Chromatogram analyzer
US20080206883A1 (en) * 2007-02-26 2008-08-28 Cody Laboratories, Inc. Hplc method for separation and detection of hydromorphone and related opioid pharmacophores
US20110054804A1 (en) * 2009-08-26 2011-03-03 Pfaff Hans Method of Improving the Resolution of Compounds Eluted from a Chromatography Device
US20120089344A1 (en) * 2010-10-07 2012-04-12 Wright David A Methods of Automated Spectral Peak Detection and Quantification Having Learning Mode
CN103063751A (zh) * 2011-10-20 2013-04-24 苏州赛分科技有限公司 一种测定美洛西林钠中聚合物杂质的检测方法
US20140149050A1 (en) * 2011-07-08 2014-05-29 Shimadzu Corporation Data processing system for chromatograph
US20160084807A1 (en) * 2014-09-18 2016-03-24 Dionex Corporation Automated method of calibrating a chromatography system and analysis of a sample
WO2016051357A1 (fr) * 2014-10-01 2016-04-07 Ldetek Inc. Détection d'impuretés de gaz avec un appareil de chromatographie en phase gazeuse

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6983213B2 (en) * 2003-10-20 2006-01-03 Cerno Bioscience Llc Methods for operating mass spectrometry (MS) instrument systems
US7391016B2 (en) * 2005-05-12 2008-06-24 Analiza, Inc. Method for quantitative analysis of mixtures of compounds
JP5308523B2 (ja) * 2009-06-01 2013-10-09 パナソニック株式会社 立体画像表示装置
DE102009035587B3 (de) * 2009-07-31 2011-03-24 Siemens Aktiengesellschaft Verfahren zur Filterung eines Chromatogramms
CN101819168B (zh) * 2010-01-06 2012-06-13 北京矿冶研究总院 一种光谱分析方法
CN103282770B (zh) * 2011-01-07 2015-08-12 株式会社日立高新技术 质量分析装置、分析法和校准试样
US9453825B2 (en) 2012-09-04 2016-09-27 Shimadzu Corporation Chromatogram peak purity determination apparatus
CN104730181B (zh) * 2013-12-18 2017-10-24 北京普源精电科技有限公司 谱峰终点调整方法及具有谱峰终点调整功能的色谱工作站
CN109964300B (zh) * 2016-10-07 2022-07-15 萨莫芬尼根有限责任公司 用于实时同位素识别的系统和方法
SG11201902667UA (en) * 2016-10-25 2019-05-30 Regeneron Pharma Methods and systems for chromatography data analysis

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4468742A (en) * 1981-03-17 1984-08-28 The Regents Of University Of California Microprocessor system for quantitative chromatographic data analysis
US5121443A (en) * 1989-04-25 1992-06-09 Spectra-Physics, Inc. Neural net system for analyzing chromatographic peaks
US6438499B1 (en) 1998-09-10 2002-08-20 Tosoh Corporation Chromatogram analyzer
US20080206883A1 (en) * 2007-02-26 2008-08-28 Cody Laboratories, Inc. Hplc method for separation and detection of hydromorphone and related opioid pharmacophores
US20110054804A1 (en) * 2009-08-26 2011-03-03 Pfaff Hans Method of Improving the Resolution of Compounds Eluted from a Chromatography Device
US20120089344A1 (en) * 2010-10-07 2012-04-12 Wright David A Methods of Automated Spectral Peak Detection and Quantification Having Learning Mode
US20140149050A1 (en) * 2011-07-08 2014-05-29 Shimadzu Corporation Data processing system for chromatograph
CN103063751A (zh) * 2011-10-20 2013-04-24 苏州赛分科技有限公司 一种测定美洛西林钠中聚合物杂质的检测方法
US20160084807A1 (en) * 2014-09-18 2016-03-24 Dionex Corporation Automated method of calibrating a chromatography system and analysis of a sample
WO2016051357A1 (fr) * 2014-10-01 2016-04-07 Ldetek Inc. Détection d'impuretés de gaz avec un appareil de chromatographie en phase gazeuse

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3545296A4

Also Published As

Publication number Publication date
US20180149629A1 (en) 2018-05-31
EP3545296A1 (fr) 2019-10-02
US10634653B2 (en) 2020-04-28
CN110234990B (zh) 2022-11-29
EP3545296A4 (fr) 2020-07-08
CN110234990A (zh) 2019-09-13

Similar Documents

Publication Publication Date Title
US10634653B2 (en) Systems, methods and devices for width-based analysis of peak traces
Ravisankar et al. A review on step-by-step analytical method validation
Bailey et al. Chemometric resolution and quantification of four-way data arising from comprehensive 2D-LC-DAD analysis of human urine
Fraga et al. Enhancing the limit of detection for comprehensive two‐dimensional gas chromatography (GC× GC) using bilinear chemometric analysis
US10605793B2 (en) Automated method of calibrating a chromatography system and analysis of a sample
US10935526B2 (en) Liquid chromatograph and method for correcting detector output value fluctuation of liquid chromatograph
Yu et al. A sensitive and accurate method for the determination of perfluoroalkyl and polyfluoroalkyl substances in human serum using a high performance liquid chromatography-online solid phase extraction-tandem mass spectrometry
Djurdjevic et al. Chemometric optimization of a RP‐HPLC method for the simultaneous analysis of abacavir, lamivudine, and zidovudine in tablets
Laursen et al. Enhanced monitoring of biopharmaceutical product purity using liquid chromatography–mass spectrometry
Brinkmann et al. Non-linear calibration functions in ion chromatography with suppressed conductivity detection using hydroxide eluents
Deshmukh et al. A review on analytical method development and validation by high performance liquid chromatography technique
US9739756B2 (en) Data processing system and method for chromatograph
Hansen Quantitative and qualitative chromatographic analysis
Birajdar New Method Development by HPLC and Validation as per ICH Guidelines
Rutan et al. Factor analysis and Kalman filter studies of severely overlapped amino acid derivatives in thin-layer chromatography
Taleuzzaman et al. Bio-Analytical Method Validation-A Review
JP6879058B2 (ja) 液体クロマトグラフ分析方法
Johnson et al. Non-target analysis of vapor mixtures using silicon nanowire array sampling and thermal desorption
Holı́k et al. Determining enantiomeric excess from overlapping HPLC peaks: discussion of errors in the methods
Rodriguez et al. Application of capillary ion chromatography and capillary ion chromatography coupled with mass spectrometry to determine methanesulfonate and inorganic anions in microliter sample volumes of Antarctic snow and ice
Sudheeshna et al. PERSPECTIVES ON ANALYTICAL METHOD VALIDATION-STRATEGIES AND IMPORTANCE IN PHARMACEUTICAL RESEARCH
Dharuman ACCURACYTARGETTED HPLCMETHODVALIDATION INVOLVINGMULTICOMPONENTDOSAGEFORMS A PROTOCOL ENRICHMENTPERSPECTIVE
Drolc Measurement of Concentration of Nitrate in Drinking Water
Jadhav Development and validation of chiral analytical methods for separtion of enantiomers of bulk drugs by hplc
Hu et al. Using singular value ratio for resolving peaks in HPLC-DAD data sets

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17874213

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2017874213

Country of ref document: EP